It would be nice if Biff weren't dependent on various ad-hoc scripts (some in bash, others in babashka), since it means newcomers to Biff (and especially newcomers to Clojure in general) have to install a variety of tools without clear purpose.<p>Luminus, for instance, is able to handle project creation, development, and deployment all within Leiningen (a build tool) itself. All you need to get started is to install Leiningen (a single command to install on any popular package manager) then run<p><pre><code> lein new luminus my-app
cd my-app
lein run
</code></pre>
No Windows-specific quirks w.r.t. live reload or anything else. Yes, it sucks that you have to install a build tool, but it's still a much simpler experience IMO. Unfortunately, Luminus feels less like a framework and more like a template web app with a bunch of assumptions about the backend and frontend baked in for you. So it's not really better than Biff either when it comes to being newbie friendly or feeling like Rails / Laravel / Django.<p>I visited the "Getting Started" section of Biff and I am immediately met with "install Babashka then run this random script off the internet". It doesn't help that they already use tons of Clojure-specific terminology like "jacking in" and assume you are using CIDER (hence Emacs). While I use Emacs and Clojure, this documentation reads like it's intended for Emacs users with years of experience in Clojure. In my opinion, this contradicts the design philosophy of "what would be helpful for ... those who are relatively new to Clojure."<p>I'll admit the plugin architecture of Biff is really nice, and I welcome decisions like using HTMX. However, there is also questionable design decisions like choosing XTDB on behalf of the user as opposed to something like SQLite or PostgreSQL, which would drastically lower the mental overhead for Clojure newcomers and even Clojure devs themselves. Although this is anecdotal evidence, most people I talked with at Clojure/conj 2023 use PostgreSQL and aren't too familiar with Datomic or even XTDB. (The fact XTDB 2.0 emphasizes SQL support says enough about the situation, I think). Out of dozens of people I talked with, I met exactly one person who actually uses Datomic in production. Nobody I met used XTDB.<p>I could go on, but I hope I've gotten my point across about three things;<p>1. as someone with several years of professional experience using Clojure, there is seemingly lots of disjoint, moving parts involved<p>2. like existing Clojure "frameworks", it still fails to feel like a Rails / Django / Laravel and feels more like "here is a project template with a bunch of decisions baked in for you". Some decisions are awesome (HTMX instead of a complicated CLJS frontend) and others are questionable (XTDB and Babashka instead of simply using PostgreSQL and Bash scripts).<p>3. the documentation feels like it's written for experienced Clojure devs, not a general audience